1. Field of the Invention
The present invention relates to a watermark generating circuit and a related method thereof, and particularly relates to a watermark generating circuit for a video encoder and a related method thereof.
2. Description of the Prior Art
FIG. 1 is a block diagram illustrating a prior art video encoder 100. The video encoder 100 includes a DCT (discrete cosine transform) device 102, a motion estimator and compensator 104, a quantizer 106, a VLE (variable length encoder) 108 and a rate controller 110. The video encoder 100 utilizes a DCT device 102 and a quantizer 106 to estimate spatial redundancy information of a video frame, and utilizes the compensator 104 to estimate time redundancy information to compress video data. Finally, the data is encoded by the VLE 108 and transmitted to a system multiplexer (not illustrated), and outputs data via a transport stream or program stream following the MPEG2 standard.
In order to make sure that the image is correct, and not destroyed or distorted in transmission, the video encoder always adds a watermark for identifying data when it is encoded.
A digital watermark technique is a technique that embeds an invisible digital watermark to an image for protecting image or copyright, and can include three stages: digital watermark generation, digital watermark embedding stage, and extraction and verification of digital watermark stage. In the digital watermark generation stage, the user generates a digital watermark, which is always a binary image, or a random value that has an equivalent value of 0 and has normal distribution. In the digital watermark embedding stage, the user utilizes an embedding key to embed the digital watermark to an image to be protected. In the extraction and verification of digital watermark stage, the user utilizes an extraction key to extract the digital watermark from the image to be detected, and compare with another digital watermark. The comparison method and results are varied according to the utilization purpose.
Digital watermarks can be classified into robust type and fragile type. The robust type watermark can be embedded to images via specific techniques, and the existence thereof can be detected even after image processing, such that the copyright can be protected. However, a fragile type watermark embedded to an image will be destroyed if the image concept is varied, such that the completeness of the image can be identified via observing if the extracted watermark is complete. Also, the fragile type watermark has a characteristic that it can locate the destroyed region of the image. The image identification technique due to digital watermark can utilize an information hiding technique to embed the digital watermark into the frequency domain.
The above mentioned watermark technique can only utilize a specific key that is easily detected. Also, most watermark techniques can only be utilized to detect I frames but cannot be utilized to detect P frames or B frames. Also, many watermark techniques utilize complicated operation structures such as Lagrangian cost function or encoding methods such as a two pass encoding method, and thus are not convenient for hardware circuit implementation and real time encoding.